輔測定錨法與心理計量法之效度評估：以政治功效感的測量爲例

A Validity Assessment of the Anchoring Vignettes vs. Psychometric Approach: The Case of Political Efficacy Measurement

張書憲(Shu-Hsien Chang) ； 黃旻華(Min-Hua Huang)

《東吳政治學報》 26卷3期 (2008/09) Pp. 115-147

https://doi.org/10.6418/SJPS.200809.0115

摘要

在民意調查學界，長久以來「項目功能差異」(differential item functioning)的問題一直困擾研究者對於心理測量效度的信任，主要的問題出在受訪者對於問卷題目和量表的感受及回應，不但存在著「人際不可比較性」(interpersonal incomparability)的問題，同時對於題目本身是否能夠真正抓住研究者所欲測量的心理態度，往往令人感到懷疑。對於上述問題，心理及教育學界主流作法是以精巧的心理計量模型來克服測量上的問題，但在2004年的《美國政治科學論叢》(American Political Science Review)中Gary King等學者提出了「輔測定錨法」(anchoring vignettes)的測量方法，宣稱此法可以徹底解決「項目功能差異」的問題。為了評估兩法的相對優缺點，本文藉由三個大學系所671位學生為受試者，進行政治功效感的心理測驗，並以上述樣本進行兩種心理測量法的效度評估。結果發現，輔測定錨法在測量效度上並不比心理計量法為佳；更甚者，輔測定錨法存在嚴重的方法論問題，使得作者對於此法的實施抱持著相當保留的態度。

關鍵字

輔測定錨法 ； 心理計量學 ； 項目功能差異 ； 人際不可比較性

並列摘要

For a long time, the problem of differential item functioning (DIF) has troubled researchers' trust in the validity of psychological measurements in the academic community of public opinion research. The main issue is not only reflected in the phenomenon that different respondents all have their own perceptions and reactions to the questionnaire and response categories, which is called ”the problem of interpersonal comparability,” but also related to the serious doubt of whether the questionnaire can tap into the attitudes that researchers want to measure.
To cope with the above issues, mainstream scholars of the psychological and educational academic community apply sophisticated models to deal with the measurement problem. However, Gary King and his colleagues in their 2004 APSR article proposed the approach of anchoring vignettes and claimed that it can solve the DIF problem thoroughly. This article tries to evaluate the measurement validity of the two approaches with an experimental survey of measuring the political efficacy of 671 enrolled students from three universities. The result indicates that the approach of anchoring vignettes is not superior to the approach of psychometrics at all in terms of validity assessment. Furthermore, anchoring vignettes suffers serious methodological problems which lead the authors to conclude with a considerable reservation on the application of anchoring vignettes to replace the psychometric measurements.

並列關鍵字

anchoring vignettes ； psychometrics ； differential item functioning ； interpersonal incomparability

參考文獻

Brady, Henry E.(1989).Factor and Ideal Point Analysis for Interpersonally Incomparable Data.Psychometrika.54(2),181-202.

Cronbach, Lee J.,Paul E. Meehl.(1955).Construct Validity in Psychological Tests.Psychological Bulletin.52(4),281-302.

Du Toit, M.(2003).IRT from SSI. BILOG-MG, MULTILOG, PARSCALE, TESTFACT.Lincolnwood, I.L.:Scientific Software International.

Embretson, Susan E.,Steven P. Reise.(2000).Item Response Theory for Psychologists.Mawwah, N. J.:Lawrence Erlbaum.

Fink, Arlene.(1995).How to Ask Survey Questions.Thousand Oaks, CA:SAGE.